Delay

Unanswered Question

Hi

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Reply from10.10.10.250: bytes=32 time<27ms TTL=255

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Reply from10.10.10.250: bytes=32 time<27ms TTL=255

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Reply from10.10.10.250: bytes=32 time=17ms TTL=255

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Reply from10.10.10.250: bytes=32 time<20ms TTL=255

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Reply from10.10.10.250: bytes=32 time<1ms TTL=255

Pls Can you Explain this for me why i have delay in the same Rreplys?and what is the reasons for this?where the Problems

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (2 ratings)
Loading.
Joseph W. Doherty Thu, 01/31/2008 - 10:20

Without any details, can only suggest two common reasons for variable ping delays.

First, transient congestion can cause data to queue, making for additional latency.

Second, hosts don't have to treat replying to a ping request as high priority. They respond when there's nothing that they consider more import to service first.

PS:

On later Cisco routers they support SLA functions that can provider a highly accurate network latency time.

Joseph W. Doherty Thu, 01/31/2008 - 10:55

For an example of the first, consider 3 hosts connected to a switch and they all transmit data at the same time. Assume the uplink from the switch has the same capacity as the individual hosts. I.e. all links are 100 Mbps.

So when the 3 hosts transmits at 100 Mbps, there's now 300 Mbps that enters the switch but only 100 Mbps can exit the switch. What the switch does is queue or buffer 2 of the 3 hosts until if finishes sending the data from 1 of the 3. While the latter 2 wait, they are delayed and you will see a jump in ping times if your ping packet also is placed in the queue or buffer.

As an example of the second condition, you're 4th host on the same switch and ping it. In the above example it's busy dealing with the data from the 3 other hosts, instead of delaying procesing their data (it's primary purpose), it responds to your ping when it has finished with those 3. You will then see a jump in ping latency because the switch was slow to respond.

suelange Thu, 01/31/2008 - 11:35

I can't tell from your post what pathway you would be using to get these results.

But you can generally find congestion easiest by looking at the show interface command for the ports and interfaces invovled and look for errors dropped packets and so on. They don't always indicate conjestion but it's a good bet. Frame network can indicate conjestion by the FECN and BECN count.

Conjestion usually results when your network is overloaded. Consider segmenting physically or with vlans.

We also learned that if the device you are pinging is a switch vlan interface, you will see that kind of up/down response time because, as the prior answer indicated, the switch doesn't consider responding to pings as a high priority. If the target address of your pings is a switch, trying pinging a workstation or other node just beyond it to see if you don't get a more stable response time.

Actions

This Discussion